On Thu, Aug 02, 2012 at 03:13:12PM -0700, Thomas DiModica wrote: > I get what you're saying: I'm confusing the semantics of how cancellation > is HANDLED with the semantics of how it is SIGNALED. The signaling > semantics are the same.
That's it. > You mean like a PTHREAD_CANCEL_GNU? It would be doable. I would personally name it PTHREAD_CANCEL_HURD to avoid increasing the already existing confusion. > If that were the case, we could merge hurd_condition_wait into > phread_cond_wait as a special case. Maybe: streamio > intermingles the two, pflocal did. > That would also handle hurd_check_cancel. Why not. -- Richard Braun
